NUCLEOPHILIC SELECTIVITY IN ATTACK AT AMIDE BONDS
Nucleophilic Slelectivityin Attack at Amide Bonds. Reactivity of Oxygen and Nitrogen Nucleophiles with N-Acetyldehydrophenylalanyl-L-proline Diketopiperazinet Thomas C. Bruice* and Diane M. McMahon
ABSTRACT :
The N-acetylated diketopiperazine (N-acetyldehydrophenylalanyl-L-proline diketopiperazine (I)) undergoes nucleophilic attack to yield the N-acetylated dipeptide (Nacetyldehydrophenylalanylproline (11)) or deacylated diketopiperazine (dehydrophenylalanylproline diketopiperazine (111)). The following specificities for position of bond cleavage have been found. (1) Primary and secondary amines, hy= 1.23) prodrazine, ammonia, and imidazole (krHzo/krDZ0 vide I11 via unassisted nucleophilic attack (kNINuc:JII]).(2) Ammonia, methoxylamine, and semicarbazide yield both I1
I
n an early synthetic study, Bergmann and Tietzman (1944) found that the reaction of N-acetyldehydrophenylalanyl-Lproline diketopiperazine (I) with glycine ethyl ester and sodium glycinate gave exclusively dehydrophenylalanylproline diketopiperazine (111) whereas in the reaction with hydroxide ion only N-acetyldehydrophenylalanylproline(11) was formed.
I1 (1)
I 0 111
It occurred to us that perhaps oxygen and nitrogen nucleophiles might exhibit nucleophilic specificities toward the endoand exoamide linkages of I, respectively. To a certain degree this is so and there is described herein studies of the competition of various nucleophiles for attack at the acetyl and lactam carbonyl carbons of I. Experimental Section N-Acetyldehydrophenylalanyl-L-prolinediketopiperazine (I) was prepared from N-acetyldehydrophenyh-proline (11) by refluxing in acetic anhydride according to the method of ~~
-f From the Department of Chemistry, University of California at Santa Barbara, California 93106. Received October 28, 1971. This research was supported from a grant from the National Institutes of Health.
and I11 in reactions requiring general base and general acid assistance to nucleophilic attack [(kgb[Nuc:J2 k,,[Nuc:]. [NucH]) [I]]. (3) The oxygen bases (formate, acetate (k?*'/ krDZ0= 2.23), phosphate, trifluoroethoxide, hydroxide, as well as Tris) provide I1 via paths of general base catalyzed attack of water (kJweak base][H20][I]) or direct nucleophilic attack. These observations are discussed in terms of the relative leaving tendencies in formation of I1 or I11 from I, ring strain in the diketopiperazine, and intramolecular participation of imide anion in ester hydrolysis.
+
Bergmann and Tietzman (1944). The material melted at 170-173' (lit. mp 172-173'). Anal. Calcd for C&1&2O3: C, 67.6; H, 5.63; N, 9.83. Found: C, 67.0; H, 5.92; N, 10.05. Dehydrophenylalanyl-L-proline diketopiperazine (111) was prepared from I and glycine (Bergmann and Tietzman, 1944). The compound was also obtained by dissolving I in acetonitrile, adding an imidazole buffer solution (0.1 M, pH 7.04) and allowing the solution to stir for approximately 30 min. The product was then precipitated by addition of a solution of NaCl, collected by filtration, and dried in vacuo over P205 (mp 170-172, lit. mp 171-172). Anal. Calcd for C I ~ H ~ ~ N ~ O ~ : C, 69.40; H, 5.83; N, 11.56. Found: C, 69.24; H, 6.07; N, 11.16. Dehydrophenylalanylproline was prepared from 111 by alkaline hydrolysis (Bergmann and Tietzman, 1944), mp 168-170 (lit. mp 168-171) (after one recrystallization from acetone and water). Anal. Calcd: C, 64.60; H, 6.20; N, 10.76. Found: C, 64.27; H, 6.36; N, 10.52. N-Acetyldehydrophenylalanyl-L-proline(11) (mp 143-145, lit. mp 140-142) was prepared according to the procedure of Behrens et al. (1940). 4-Benzylidene-2-methyI-2-oxazoline-5-onewas prepared from a-N-acetimidocinnamic acid according to the method of Darkin and West (1928): mp 152.5-153", lit. mp 152; infrared (ir) spectrum carbonyl 1780 cm-', no NH stretch observed; nuclear magnetic resonance (nmr) spectrum (in 7 values) 1.80-2.67 ( 5 ) , 2.68 (l), 7.6 (3) (CDC13). N-Acetyldehydrophenylalanyl-L-prolinamide(IV) was prepared by treatment of I with aqueous ammonia according to the method of Bergmann and Tietzman (1944): mp 234.5235', lit. mp 239-241 '. pK, Determinations. The pK, of dehydrophenylalany1-Lproline diketopiperazine (111) was determined at 30', p = 1 .O (with KCl), by measurement of the optical density of the anion (at A, 322 nm) at various pH values. The theoretical sigmoidal curve giving the best fit to the experimental data resulted in a pK, of 10.77. The pK, of I1 was calculated on the basis of eq 2 and 3. At the wavelength employed (322 nm) only the dianion of I1 abVO BIOCHEMISTRY , L. 11,
NO.
7, 1 9 7 2
1273
BRUICE
AND MCMAHON
[I0] and (2) the fraction of product arising from hydroxide ion catalyzed hydrolysis can be calculated from the observed pseudo-first-order rate constant ( k o b s d ) , which has kinetic terms both in ammonia and hydroxide ion. (3) sorbed. Since the concentration of I1 was constant throughout the experiment, it was assumed that OD-1 = OD,Z (OD, is the absorption at 322 nm of I1 at pH >> pK,). From eq 2 and 3, eq 4 is derived.
Results The reaction of I with nineteen nucleophilic species has been investigated under the pseudo-first-order conditions of [N:] [NH] = [NT] > [I0](where N : represents nucleophile free base and NH its conjugate acid). The experimental conditions and derived second-order rate constants for reaction of N : with I are presented in Table I. In this study it was found that, in general, primary andsecondary alkylamines yield (111), whereas the oxygen nucleophiles studied (acetate, formate, hydroxide, and phosphate) give 11. The deuterium solvent kinetic isotope effects were determined as k H i O / k D z O = 2.23 for the hydrolysis catalyzed by acetate and k H z O / k D 2 0 = 1.23 for imidazole. The Bronsted /3 values found for nucleophilic attack of amines is 0.92, whereas that for the catalysis of hydrolysis by the oxygen bases is 0.62 (Figure 1). Spectrophotometric studies with 2,2,2-trifluoroethoxide, showed that both I11 and I1 are formed. The overall rate of appearance of I1 in the presence of ~ 0 . M1 trifluoroethanol is 1/20ththat anticipated from kHo[HO-] at pH 11.5. Studies with ethanol at pH 10.35 show no such similar depression of the hydroxide rate with increasing ethanol concentration. Work done in neat trifluoroethanol (1.16 N in trifluoroeth277 nm) is very oxide) shows a product whose spectrum (A,, similar to that of I1,,A(, 275 nm). The diketopiperazine, 111, has a A,, of 292 nm in the same trifluoroethanol solution, thereby excluding it as the product. The ir spectrum of reaction products formed on a preparative scale in neat trifluoroethanol shows the appearance of a new carbonyl stretching band at 1760 cm-1 and no typical carboxylic acid absorption in the 2600- to 3300-cm-l range in agreement with formation of an ester of 11. Spectrophotometric studies of the reaction of I at constant pH in the presence of ammonia showed that three products I1 (cia HO--catalyzed hydrolysis), 111, and N-acetyldehydrophenylalanyl-L-prolinamide (IV) are formed, Buffer dilution studies established the disappearance of I to be both first and second order in total ammonia buffer concentration, [ N H 3 ~ Product ]. analysis of the reaction mixture showed that formation of 111 is proportional to both a first- and secondorder kinetic term in ammonia concentration, whereas formation of IV is proportional only to a second order term in ["IT].
+
The pK, of I1 is an average, calculated employing eq 4 and experimental values of OD32znrnfound at several values of a H . This method was employed because a value for the extinction coefficient of the anion was unobtainable due to the high pK, of the amide. The pK, value obtained was 13.9 i 0.2. The pK, values of glycylglycine and 2,2,2-triRuoroethylamine at temperatures above 30" were determined by means of half-neutralization employing a Metrohm type H hightemperature glass electrode and a Radiometer 26 pH meter used in conjunction with a constant temperature cell previously described by Bruice and Marquardt (1962). Prior to measurements, the glass electrode was calibrated at each temperature against standard buffer. Kinetic Measurements. The rates of hydrolysis of I at 30 i 0.1 O (p = 1.O) in water were followed by measuring the appearance of I1 at 278 nm or I11 at 292 nm (320 nm for anion) with a Gilford 2000 recording spectrophotometer. In all spectrophotometric work a stock solution of I in acetonitrile was used. For the spectrophotometric rates run on the Gilford 2000, an aliquot (20-30 pl) of stock solution of I was added to 2.5-3.0 ml of buffered solutions. The spectra of the solutions upon completion of the reaction were identical with those of I1 or I11 in the appropriate buffer solution. Rate constants for hydroxide ion catalyzed hydrolysis were determined titrimetrically ( p = 1.0 with KCI) at constant pH with a Radiometer TTT l b autotitrator and followed spectrophotometrically with a Cary 15, RS. Pseudo-first-order rate constants were calculated with an Olivetti Underwood Programma 101 program for (log [ ( A !- A o ) / ( A o- A,)] 1;s. time) infinity or Guggenheim-type plots, employing least-squares analysis for slope and intercept. All pH's were determined (at 30.0 i 0.1') with a Radiometer Model 22 pH meter equipped with a Model 630 scale expander and a combined glass calomel electrode (Radiometer GK 2021C). The ApD correction (Fife and Bruice, 1961) was employed for work done in DzO. Product Analysis. Extinction coefficients of 11, 111, and Nacetyldehydrophenylalanyl-L-prolinamide (IV) were determined at 278, 282, and 292 nm, respectively. Stock solutions of the compounds in CH3CN were pipetted in microliter quantities into a cuvet holding 3.0 ml of solution. The optical density was measured for each compound at at least six concentrations at each wavelength. Product analyses for the aminolysis were done at a constant concentration of I. Final values were obtained by solution of the simultaneous equa[IIII [I11 = tions (eq 5 and 6) and the fact that (1) [IVI
+
+
(9) leading to
A plot of [III]/[IV]
L'S.
l/[NHBT]was found to be linear (plot
NUCLEOPHILIC
SELECTIVITY IN ATTACK AT AMIDE BONDS ~
TABLE I : Rate Constants
of Nucleophiles in the Reaction of I11 (p = 1.O, T = 30.0').
Nucleophile
PK.
1. HP042. -OH 3. CHICOO4. HCOO5. Tris 6. Gly-Gly' 7. Imidazole. 8. Morpholinea 9. Ethylenediamine
6.7 15.6 4.74 3.6 8.10 8.17 6.97 8.3 7.05. 10.1 5.60 7.65 9.48 4.69 9.23 3.65 6.17 10.6 7.7
10. CF3CH2NHza 11. NHZNHZ0 12. HOCHzCHzNHz 13. CHzONHz 14. "3" 15. Semicarbazide 16. Triethylenediamine 17. tert-Butylamine 18. N-Ethylmorpholine a
p =
Concn (M) Range
pH Range
0.214-0.680
6.93-5.97 9.03-10.35 4.74 3.53 8.05 8.03-8.89 7.0-7.2 8.48 7.76-7.76
0.2-1.0 0.36-1.0 0.02-0.10 0.02-0.10 0.1-0.5 0.02-0.10 0.01-0.15 0.12-0.25 0.03-0.07 0.06-0.23 0.2-1 .o 0,055-0.330 0.3-1.0 0 . 4 1 .o 0.14-0.48 0.2-1 .o
NO. Of kobsd ki Detmd (1 mole-' min-1) 8 6 4 3 4 12 10 4 20
5.55 7.75 9.00 3.88-4.81 9.56 3.88 6.17 10.60 7.7
0.043 6080 0.00504 0,0019 0.762 8.51 6.52 3.70 0.676 741 .O 0.0466 318.0 328 0.318 37.1 0.032 O.OOO63 b,c b,c
4 4 4 16 5 4 3 4 4
0.5. * No catalysis of hydrolysis found. c Compound I1 formed as product.
not shown), and of slope kn/krn2and intercept kn2/k',2. From the kinetic eq 11, which relates the pseudo-first-order rate
constant (kabsd)for disappearance of I to the concentrations of HO- and N&T, eq 12 is derived. From eq 12, a plot not
shown Of (kobsd - ~HO[HO-])/["~T] US. ["IT] i S Of slope (k'+ k,z) and intercept k,. From the values of k,/k',~ obtained from the plot of eq 10 and the values for krn2 knz and the k, obtained from a plot of eq 12, the apparent rate constants based on total [NH3] [NH4+] at pH 9.56 were determined to be approximately k, = 12.9 min-I M - ~ , kn2 = 22,6min-'~-~,andk',2 = 3 4 . 3 m i n - ' ~ - ~ . a-Effect Compounds. The reaction of hydrazine with I gives I11 as the only product whereas methoxylamine and semicarbazide yield both I11 and the corresponding amide of 11. For methoxylamine, third-order rate constants were obtained from plots of k,bsd/[amide~]us. [amideITat several pH values. The termolecular term for methoxyaminolysis is composed of both general acid and general base assisted nucleophilic terms, such that
+
From eq 14 a plot of the left-hand side of the equation us. UH is of slope k,.K, and intercept kgbKaZ(Figure 2). The calculated specific constants were found to be kgb = 2.65 min-l ~ - 2and k,, = 3.08 min-I M-2. Product analysis of the methoxylamine reaction carried out at pH 6.5 ( p = 1.0; pH maintained constant with a pH-Stat) established the amide to be the major product (Amx (final) 281 nm). Products of methoxylaminolysis at pH 4.7 contained a larger amount of I11 than that present at pH 6.5 (Aux (final) 268 nm). These results suggest that kgb is associated more with nucleophilic attack at the lactam carbonyl and k,. at the N-acetylcarbonyl group, respectively. This is born out by comparison of product analysis and kinetic terms. If one assumes that unassisted
+
+
8.0
6.0
4.0 r 0 rn
-
2.0
4.0
P
I
I
6.0
8.0
10.0
I
12.0
I
14.0
pK,
From eq 13 it follows that
1 : Plot of pK, of conjugate acid of nucleophilic species us. the log of the calculated second-order rate constant ( M - ~min-') for reaction of nucleophile with N-acetyldehydrophenylalanyl-LFIGURE
proline diketopiperazine. BIOCHEMISTRY,
VOL.
11,
NO.
7, 1 9 7 2
1275
BRUICE AND MCMAHON
aH
x 105M
Plot of the apparent termolecular rate constants [k,20bsd] times the sum squared of the dissociation constant of the conjugate acid of methoxylamine plus the hydrogen ion activity [(K, aE)2] us. hydrogen ion activity. From the slope and intercept the values of the termolecular constants for self-assisted general acid and general base catalysis ( k g b ) may be derived. catalysis
FIGURE 2:
follows. (1) Primary and secondary amines, hydrazine, ammonia, and imidazole react with I Diu unassisted nucleophilic attack (k~[N:l)at the N-acetyl carbonyl to yield 111. (2) Ammonia, methoxylamine, and semicarbazide exhibit in their reaction with I, terms second order in buffer indicating general base (k,b[N:l2) and/or general acid (kg,[N:][NHl) catalysis of nucleophilic attack and yield both 11 and 111.(3) The oxygen bases (formate, acetate, phosphate, trifluoroethoxide, hydroxide, as well as Tris) react with I to yield the dipeptide 11. In the case of trifluoroethoxide I11 is also obtained. The deuterium kinetic solvent isotope effect (knH'O/knD'O) was determined to be 2.23 for the reaction of acetate with I. (4) Trifluoroethoxide but not ethanol greatly depresses the rate constant (kEo[HO-]) for hydrolysis of I. The selectivity of bases toward the endo- and exocyclic imide bonds of I to yield I1 or 111, respectively, may be explained if one accepts the following concepts : (i) nucleophilic attack to yield I11 is favored by the lower basicity of the leaving group and the fact that production of an acyl derivative of I1 is difficult due to the retrograde intramolecular displacement by imide anion to regenerate I (V; intramolecular attack
+
Ac-N
c=o nucleophilic attack (k,) of methoxylamine yields only I11 (as in the case of all other amines), then it can be shown at pH 6.5 where only free methoxylamine is present that
One explanation of the fact that kgb [CH30NH2]2/k,[CH3ONH2] > [amide]/[III] is that the free amine assistance of nucleophilic attack is also occurring (as in the case of NH3) at the acetyl position, such that
+
where kgb determined kinetically equals k,b"' kgbamlde; the amine assisted general base catalysis of attack of amine to provide I11 and the amide of I1 as products. Tertiary Amines. Hydrolysis of I was catalyzed by imidazole but not by the tertiary alkyl amines N-ethylmorpholine and only very slightly by triethylenediamine. The product of the reaction in both N-ethylmorpholine and triethylenediamine buffers was 11, whereas the product of the imidazolecatalyzed reaction was 111. Discussion
N-Acetyldehydrophenylalanyl-L-proline diketopiperazine I possesses two imide bonds which are susceptible to nucleophilic attack; these are an acetyl imide bond whose scission yields dehydrophenylalanylproline diketopiperazine 111 and an endocyclic imide bond whose hydrolytic scission yields the N-acetyl dipeptide, N-acetyldehydrophenylalanylproline (11). Our purpose has been to determine the selectivity of nucleophiles (if any) in attacking the endo- and exocyclic imide bonds and to provide a mechanistic rationale for this selectivity. Our experimental findings may be summarized as
1276
B I O C H E M I S T R Y , VOL.
11,
NO.
7, 1 9 7 2
L V
of imide anion upon the ester bond is a well-studied reaction (Bruice and Benkovic, 1966a,b); and (ii) formation of I1 is favored due to release of ring strain (the greater lability of diketopiperazines to hydrolysis, as compared to peptides, has been suggested to be due to ring strain (Corey, 1938; Edwards and Meacock, 1958; Sykes et al., 1966) so that when leaving tendency is of little consequence both I1 and I11 are favored. Also, if both leaving tendency and retrograde imide anion displacement are unimportant then relief of ring strain remains and I1 is the sole product. In what follows, or for that matter in any study of competing processes, it must be understood that reaction paths are identified by product analysis and that inability to detect a few percent or more of a minor product entails the assumption that the reaction path is not present. Thus, if substrate passes through path A to yield 90% of A and through path B to yield an undetectable 10% of B the assumption is made that the reaction i s specific for A. Obviously, for this case, AFB exceeds AFA by but 1.3 kcal/mole-l which hardly falls into the range of high selectivity. More importantly, this feature assures us that within a series of like nucleophiles competing for two positions on a substrate the entire series may attack at one position but in a few instances attack may occur at both positions. To rationalize these exceptions may require the explanation of physical features which must account for a difference of as little as a kilocalorie or less in free energies of activation. To attempt to do so is often unrealistic. The same considerations may apply to competing paths for attack at a single position which differ in kinetic order (i.e., k,[N:] us. kgb[N:I2).Therefore, with some trepidation we proceed. The pK. values of the N-H bonds of I11 and I1 have been determined to be 10.8 and 13.9, respectively. Thus, 111- is a better leaving group than is 11-. Consideration of this feature
*
*
NUCLEOPHILIC SELECTIVITY IN ATTACK AT AMIDE BONDS
SCHEME I
111-
11-
pK’a of formation 10.8
13.9
alone dictates that nucleophilic attack should be favored at the acetyl carbonyl carbon which is as observed in simple nucleophilic attack of the nitrogen bases. As the difference in the pK, values of the conjugate acids of the amine nucleophile and leaving group become greater it is anticipated that displacement of the leaving group by amine nucelophile should become most difficult and that general catalysis of nucleophilic attack should come into play. This expectation is based on generally known and widely appreciated experimental results (Bruice and Benkovic, 1964; Kirsch and Jencks, 1964) and is amplified by the fact that the unassisted aminolysis reaction is quite sensitive to the basicity of the amine, as shown by a BrQnsted /3 value of 0.92 (Figure 1). In addition, it is known that the a-effect provides a greater rate enhancement to the general catalytic terms kgb and k,, than to the nucleophilic term k, for amines (Bruice and Benkovic, 1964; Fedor and Bruice, 1964). It is to be expected, therefore, that k,b and k,, paths should favorably compete with the k , path for a-effect amines of low pK, as seen for methoxylamine and semicarbazide. Unlike the aminolysis of acetylimidazole free base (Jencks and Gilchrist, 1968 ; Fersht and Jencks, 1970), there is no evidence for general catalysis in reaction of primary and secondary amines with I to yield 111. This may be attributed to the pK, of the leaving groups in the case of acetylimidazole (14.2) (Bruice and Herz, 1964) and I (10.8). Since in general catalysis, protonation of the leaving group or proton abstraction from the attacking nucleophile is occurring in the transition state, the relative base strength of nucleophile and leaving group is of less importance and this levelling effect should lead to less selectivity. This feature explains the fact that methoxylamine, semicarbazide, and ammonia through their general catalytic paths yield with I both I11 and amides of 11. Nucleophilic attack of oxygen nucleophiles, save hydroxide ion, should be disfavored at the ring carbonyl group due to the expectation that the product would not be stable since intramolecular imide anion attack should regenerate I (V). Indeed, this appears to be the case for trifluoroethoxide ion and the decrease in the rate constant for hydroxide ion catalysis of the formation of I1 on increase in trifluoroethanol concentration finds explanation in initial attack of the alkoxide ion upon I to yield an equilibrium concentration of the ester of I1 (Scheme I). Since hydroxide ion yields from I only I1 the fact that 111is formed in the presence of trifluoroethanol can only mean that the alkoxide ion can attack at both imide bonds. This may be attributed to the fact that the pK,’ of trifluoroethanol (12.36) (Ballinger and Long, 1959) is comparable to those of the leaving groups. The greater proportion of product arises from attack at the acetyl carbonyl to expel the better leaving group. It is anticipated that general base assistance of attack of water should be the preferred mode of reaction of formate, acetate, and phosphate with I. This expectation follows from the large difference in pK, values of the conjugate acids of the oxygen bases and the leaving group in formation of either I1 or 111. General base assistance of the attack of HzO @e., kgb
I
H@ Ac
coo0
[N:I[HzOl) has been reported for carboxyl anion nucleophiles (RCOO-) to come into play in the hydrolysis of phenyl acetate when the difference in pK, of RCOOH and pehnol are such that simple nucleophilic attack ( k , [ N : ] ) is precluded (Gold et al., 1968). That the mechanism is general base assistance of water attack is borne out by the deuterium solvent kinetic isotope effect of greater than two for acetate. Similar results were obtained in the study of the hydrolysis of Nacetylimidazole, where the weak oxygen bases were found to participate as general base catalysts (Oakenfall er al., 1971; Oakenfall and Jencks, 1971). For general base assistance of the attack of water, bond formation in the transition state should be far toward product and in the nucleophilic attack of hydroxide ion the basicity of the entering nucleophile exceeds those for the departing groups. This leads to the expectation that leaving tendency may not have a great effect upon selectivity for the position of attack upon I by HO- or H 2 0 catalyzed by weak oxygen nucleophiles. That all oxygen bases bring about hydrolysis of the ring imide bond may be related to the probability of release of strain on opening the diketopiperazine ring and to the fact that intramolecular nucleophilic attack (as in Scheme I) to regenerate, though not impossible (Hegarty and Bruice, 1969, 1970), is made energetically less likely due to the requirement of attack of imide anion upon carboxyl anion. It is interesting to note that reaction of hydroxide with I is strictly first order in hydroxide ion. This feature is shared in the HO- catalysis of hydrolysis of acetylimidazole (leaving group 14.2). In contrast, p-nitroacetanilide and p-formylacetanilide hydrolysis (leaving group pK.’ = 13.8 and 14.1, respectively) exhibit terms second order in HO- dictating product to arise via dianionic tetrahedral intermediates (Eriksson and Bratt, 1967; Pollak and Bender, 1970; Bender and Thomas, 1961 ; Jencks and Salvesen, 1971). An anomaly exists in thereaction of ammonia with I whereit is found to be subject to self general catalysis though primary and secondary amines of greater and much less basicity do not exhibit general catalytic terms. One possible explanation for this finding may reside in the lessened steric requirement of ammonia. Thus, it has been well established that steric requirements are amplified in kgband k,, terms as compared to k , in the aminolysis of thiol and phenyl esters (Bruice el ai., 1967; Fedor and Bruice, 1964). That steric requirements may be particularly important in reactions at the imide bonds of I finds support in the fact that the two tertiary amines examined BIOCHEMISTRY VO , L. 11,
NO.
7, 1 9 7 2
1277
CLARK
UV-ethylmorpholine and triethylenediamine) show no and little reactivity, respectively, toward I. Added in Proof The structures of I and derived products have been drawn with the phenyl ring cis to the ring carbonyl group. Recent X-ray and nmr data of Brocklehurst e l a/. (1971) dictate that the structures should have a trans configuration. This in no way influences the results of this study. References Ballinger, P., and Long, F. A. (1959), J . Amer. Chem. SOC. 81,1050. Behrens, 0. K., Doherty, D. G., and Bergmann, M. (1940), J . Biol. Chem. 136, 61. Bender, M. L., and Thomas, R. J. (1961), J. Amer. Cheni. SOC. 93,4183. Bergmann, M., and Tietzmann, J. E. (1944), J . Biol. Chem. 154,536. Brocklehurst, K., Bywater, R. P., Palmer, R. A., and Patrick, R. (1971), Chem. Coitirmn., 632. Bruice, T. C., and Benkovic, S. J. (1964), J . Amer. Cheiii. SOC. 86,418. Bruice, T. C., and Benkovic, S. J. (1966), Bioorganic Mechanisms, Vol. I, New York, N. Y., W. A. Benjamin Co., pp 181(a), 191(b). Bruice, T. C., Donzel, A., Huffman, R. W., and Butler, A. K. (1967),J. Amer. Chem. SOC.89,2106. Bruice, T. C., and Herz, J. L. (1964), J . Amer. Cliem. SOC. 86,4109.
AND
OGILVIE
Bruice, T. C., and Marquardt, F. H. (1962), J. Amer. Cfiem. SOC.84, 365. Corey, R. B. (1938), J. Amer. Chem. SOC.60,1538. Darkin, H. D., and West, R. (1928), J. Biof. Chem. 78,745. Edwards, J. T., and Meacock, S. C. R. (1958), J. Chem. SOC., 2008. Eriksson, S. O., and Bratt, L. (1967), Acta Chem. Scund. 21, 1812. Fedor, L. R., and Bruice, T. C. (1964), J . Amer. Chem. SOC. 86,4117. Fersht, A. R., and Jencks, W. P. (1970), J. Amer. Chem. SOC. 92,5432,5442. Fife, T. H., and Bruice, T. C. (1961), J. Phys. Chem. 65, 1079. Gold, V., Oakenfall, D. E., and Riley, T. (1968), J. Chem. SOC., A , 515. Hegarty, A. F., and Bruice, T. C. (1969), J. Amer. Chem. SOC. 91,4924. Hegarty, A. F., and Bruice, T. C. (1970), J . Amer. Chem. SOC. 92, 6575. Jencks, W. P., and Gilchrist, M. (1968), J . Amer. Chem. SOC. 90,2622. Jencks, W. P., and Salvesen, K. (1971), J. Amer. Chem. SOC. 93,1419. Kirsch, J. F., and Jencks, W. P. (1964), J. Amer. Chem. SOC. 86, 833,837. Oakenfall, D. G., and Jencks, W. P. (1971), J. Amer. Chem. SOC.93,178. Oakenfall, D. G., Salvesen, K., and Jencks, W. P. (1971), J. Ainer. Chem. SOC.93,188. Pollak, R. M., and Bender, M. L. (1970), J. Amer. Chem. SOC. 92, 7190. Sykes, B. D., Robertson, E. B., Dunford, H. B., and Konasewich, D. (1966), Biochemistry 5,697.
Aspartokinase I-Homoserine Dehydrogenase I of Escherichia coli K12 A. Subunit Molecular Weight and Nicotinamide-Adenine Dinucleotide Phosphate Binding? Richard B. Clark1 and James W. Ogilvie*
The aspartokinase I-homoserine dehydrogenase I complex isolated from Escherichia coli K12 was denatured in urea and sodium dodecyl sulfate and the subunits produced were examined by dodecyl sulfate-polyacrylamide gel electrophoresis and found to be homogeneous in size with a molecular weight of 85,000. The binding of NADP+ by the enzyme was investigated at pH 8.79 in the presence of 0.01 M LABSTRACT :
T
threonine by gel filtration. The number of NADP+ binding sites per enzyme subunit was found to be 0.50 i 0.01 and the dissociation constant for the enzyme-NADP+ complex was determined to be 5.0 i 0.2 PM. The significance of these findings to the possible structure of the native aspartokinase I-homoserine dehydrogenase I complex is discussed.
he aspartokinase I-homoserine dehydrogenase I (AKIHSDI) complex isolated from Escherichia coli possesses properties which make it particularly interesting from a func-
tional as well as a structural standpoint. A number of its properties suggest that this enzyme complex may play a key role in the control of the rate of biosynthesis of some of the
t From the Department of Biochemistry, University of Virginia, School of Medicine, Charlottesville, Virginia. Receiaed June 21, 1971. Supported by research grants from the National Institutes of Health (GM 16917) and the National Science Foundatica (GB 6894). $ Taken in part from a thesis submitted by Richard Bell Clark to the
University of Virginia in partial fulfillment of the requirements for the degree of Doctor of Philosophy. Present address: Department of Pharmacology, University of Colorado, School of Medicine, Denver, Colorado.
1278
BIOCHEMISTRY,
VOL.
11,
NO.
7, 1 9 7 2
